Java ParserFactory类代码示例

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
@Override
public void doInitialize(final UimaContext aContext) throws ResourceInitializationException {
	try {
		parserChunkingModel.loadModel(ParserModel.class, getClass().getResourceAsStream("en-parser-chunking.bin"));
	} catch (final BaleenException be) {
		getMonitor().error("Unable to load OpenNLP Language Models", be);
		throw new ResourceInitializationException(be);
	}

	try {
		parser = ParserFactory.create((ParserModel) parserChunkingModel.getModel());

	} catch (final Exception e) {
		getMonitor().error("Unable to create OpenNLP parser", e);
		throw new ResourceInitializationException(e);
	}
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
@Override
public void doInitialize(final UimaContext aContext) throws ResourceInitializationException {
	try {
		parserChunkingModel.loadModel(ParserModel.class, getClass().getResourceAsStream("en_parser_chunking.bin"));
	} catch (final BaleenException be) {
		getMonitor().error("Unable to load OpenNLP Language Models", be);
		throw new ResourceInitializationException(be);
	}

	try {
		parser = ParserFactory.create((ParserModel) parserChunkingModel.getModel());

	} catch (final Exception e) {
		getMonitor().error("Unable to create OpenNLP parser", e);
		throw new ResourceInitializationException(e);
	}
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public double scoreStructure(String ca, String q, String passage, boolean verbose) throws InvalidFormatException, IOException{
	POSTaggerME parserModel = new POSTaggerME(new POSModel(new FileInputStream(new File("en-pos-model.bin"))));
	Tokenizer tokenizer = new TokenizerME(new TokenizerModel(new FileInputStream(new File("en-token.bin"))));
	Parser parser = ParserFactory.create(new ParserModel(new FileInputStream(new File("en-parser.bin"))));
	double score = 0;
	
	Parse[] questionParse = ParserTool.parseLine(q, parser, 1);
	Parse[] passageParse = ParserTool.parseLine(q, parser, 1);
	
	if (passage.contains(ca)) {
		for (int i =0; i < questionParse.length; i++) {
			score += matchChildren(questionParse[i],passageParse[i]);
		}
	}
	
	return score;
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public void parserTest1() throws IOException {
	if (!this.modelsAreInitialized) init();
	Parser parser = ParserFactory.create(
			this.parserModel,
			20, // beam size
			0.95); 
	Parse[] results = ParserTool.parseLine("Jane Austen was very modest about her own genius ."+this.q,
			parser, 1);
	Parse[] qResults = ParserTool.parseLine(this.q,parser, 1);
	Parse[] rChn = (results[0].getChildren())[0].getChildren();
	
	results[0].expandTopNode(results[0]);
	for (int i = 0; i < results.length; i++) {
		results[i].show();
	}
	for (int i = 0; i < qResults.length; i++) {
		qResults[i].show();
	}
	System.out.print("\n\n");
	for (int i = 0; i < rChn.length; i++) {
		rChn[i].show();
		System.out.print("\n");
	}
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public static Parse[] parsePassageText(String p) throws InvalidFormatException{
	
	//initialize 	 
	SentenceDetectorME sentenceDetector = new SentenceDetectorME(sentenceModel);
	Parser parser = ParserFactory.create(
			parserModel,
			20, // beam size
			0.95); // advance percentage
 	 	 
	String[] sentences = sentenceDetector.sentDetect(p);
	Parse[] results = new Parse[sentences.length];
	for (int i=0;i<sentences.length;i++){
		String[] tks = SimpleTokenizer.INSTANCE.tokenize(sentences[i]);


		String sent= StringUtils.join(tks," ");
		System.out.println("Found sentence " + sent);
		Parse[] sentResults = ParserTool.parseLine(sent,parser, 1);
		results[i]=sentResults[0];
	}
	return results;
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public Parse[] parsePassageText(String p) throws InvalidFormatException{
	if (!modelsAreInitialized)init();
	//initialize 	 
	SentenceDetectorME sentenceDetector = new SentenceDetectorME(this.sentenceModel);
	Parser parser = ParserFactory.create(
			this.parserModel,
			20, // beam size
			0.95); // advance percentage
	//find sentences, tokenize each, parse each, return top parse for each 	 	 
	String[] sentences = sentenceDetector.sentDetect(p);
	Parse[] results = new Parse[sentences.length];
	for (int i=0;i<sentences.length;i++){
		String[] tks = SimpleTokenizer.INSTANCE.tokenize(sentences[i]);
		//StringTokenizer st = new StringTokenizer(tks[i]); 
		//There are several tokenizers available. SimpleTokenizer works best

		String sent= StringUtils.join(tks," ");
		System.out.println("Found sentence " + sent);
		Parse[] sentResults = ParserTool.parseLine(sent,parser, 1);
		results[i]=sentResults[0];
	}
	return results;
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public EnglishIndexer() throws Exception {
      mDicts = new EnglishDictionaries();
      mBeamSize = ConfigurationManager.getConfiguration().getInt("BeamSize");
      InputStream modelStream = null;

      modelStream = getClass().getResourceAsStream("/opennlp15model-sa/en-sent.bin");
SentenceModel model = new SentenceModel(modelStream);
mSentenceDetector = new SentenceDetectorME(model);
modelStream.close();

modelStream = getClass().getResourceAsStream("/opennlp15model-sa/en-token.bin");
mTokenizer = new EnglishTokenizer(modelStream, mDicts);
modelStream.close();

// The parser model is about 15x the size of chunking model. 
// Keep this in mind when using Deep Parsing.
modelStream = getClass().getResourceAsStream("/opennlp15model-sa/en-pos-maxent.bin");
//POSModel posModel = POSTaggerUtils.createPOSModel(modelStream);
POSModel posModel = new POSModel(modelStream);
mTagDictionary = posModel.getTagDictionary();
mPosTagger = new POSTaggerME(posModel);
modelStream.close();

modelStream = getClass().getResourceAsStream("/opennlp15model-sa/en-chunker.bin");
ChunkerModel chunkerModel = new ChunkerModel(modelStream);
mChunker = new ChunkerME(chunkerModel);
modelStream.close();

modelStream = getClass().getResourceAsStream("/opennlp15model-sa/en-parser-chunking.bin");
ParserModel parserModel = new ParserModel(modelStream);
mParser = ParserFactory.create(parserModel);
modelStream.close();

  }
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
private void loadResource() throws InvalidFormatException, IOException {
	if (parser == null) {
		InputStream is = OpenNLPParser.class.getClassLoader().getResourceAsStream(PARSER_MODEL);
		ParserModel model = new ParserModel(is);
		parser = ParserFactory.create(model);
		is.close();
	}

}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public static Parse parse(String input) {
    ParserModel model = (ParserModel) models.get(ParserModel.class);
    Parser parser = ParserFactory.create(model);
    Parse topParses[] = ParserTool.parseLine(input, parser, tokenizer(), 1);
    return topParses[0];

}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public OpenNLP() throws Exception {
	tagger = new POSTaggerME(new POSModel(this.getClass().getClassLoader()
			.getResourceAsStream("models/en-pos-maxent.bin")));
	parser = ParserFactory.create(new ParserModel(this.getClass().getClassLoader()
			.getResourceAsStream("models/en-parser-chunking.bin")));
	tokenizer = new TokenizerME(new TokenizerModel(this.getClass().getClassLoader()
			.getResourceAsStream("models/en-token.bin")));
	stemmer = new EnglishStemmer();
	log.info("OpenNLP models loaded");
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public ParserExtractor() {
	sentenceDetector = new SentenceDetector(Consts.EN_SENT_MODEL);
	
	try (InputStream modelIn = ParserExtractor.class.getClassLoader()
			.getResourceAsStream(Consts.EN_PARSER_MODEL);){
		ParserModel model = new ParserModel(modelIn);
		parser = ParserFactory.create(model);
	} catch (IOException e) {
		e.printStackTrace();
	}
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public SentenceSimilarity() {
	File modelsDir = new File(this.modelsPath);

	this.parserMFile = new File(modelsDir, "en-parser-chunking.bin");
	this.sentDetectorMFile = new File(modelsDir, "en-sent.bin");
	this.posMFile = new File(modelsDir,"en-pos-maxent.bin");

	InputStream sentModelIn = null;
	FileInputStream parserStream;
	try {
		//for finding sentences
		sentModelIn = new FileInputStream(sentDetectorMFile);
		this.sentenceModel = new SentenceModel(sentModelIn);
		//for finding POS
		FileInputStream posModelStream = new FileInputStream(posMFile);
		POSModel model = new POSModel(posModelStream);
		this.tagger = new POSTaggerME(model);
		//for parsing
		parserStream = new FileInputStream(parserMFile);
		this.parserModel = new ParserModel(parserStream);
	} catch (FileNotFoundException e2) {
		// TODO Auto-generated catch block
		e2.printStackTrace();
	} catch (IOException e) {
		// TODO Auto-generated catch block
		e.printStackTrace();
	}
	
	sentenceDetector = new SentenceDetectorME(this.sentenceModel);
	parser = ParserFactory.create(
			this.parserModel,
			20, // beam size
			0.95); // advance percentage
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public Parse[] parsePassageText(String p) throws InvalidFormatException{
	if (!modelsAreInitialized)init();
	//initialize 	 
	SentenceDetectorME sentenceDetector = new SentenceDetectorME(this.sentenceModel);
	NameFinderME nameFinder = new NameFinderME(this.nerModel);
	Parser parser = ParserFactory.create(
			this.parserModel,
			20, // beam size
			0.95); // advance percentage
	//find sentences, tokenize each, parse each, return top parse for each 	 	 
	String[] sentences = sentenceDetector.sentDetect(p);
	Parse[] results = new Parse[sentences.length];
	for (int i=0;i<sentences.length;i++){
		//String[] tks = SimpleTokenizer.INSTANCE.tokenize(sentences[i]);
		
		//StringTokenizer st = new StringTokenizer(tks[i]); 
		//There are several tokenizers available. SimpleTokenizer works best
		Tokenizer tokenizer = SimpleTokenizer.INSTANCE;
		for (int si = 0; si < sentences.length; si++) {
	        Span[] tokenSpans = tokenizer.tokenizePos(sentences[si]);
	        String[] tokens = Span.spansToStrings(tokenSpans, sentences[si]);
	        Span[] names = nameFinder.find(tokens);
	        for (int ni = 0; ni < names.length; ni++) {
	            Span startSpan = tokenSpans[names[ni].getStart()];
	            int nameStart = startSpan.getStart();
	            Span endSpan = tokenSpans[names[ni].getEnd() - 1];
	            int nameEnd = endSpan.getEnd();
	            String name = sentences[si].substring(nameStart, nameEnd);
	            System.out.println(name);
	        }
	    }
		String sent= StringUtils.join(tokenizer," ");
		System.out.println("Found sentence " + sent);
		Parse[] sentResults = ParserTool.parseLine(sent,parser, 1);
		results[i]=sentResults[0];
	}
	return results;
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
public void loadLexParser() throws IOException {
	String model_file = prop.getProperty("parse.model");
	System.err.printf("Loading parser from %s ... ", model_file);
	parser = ParserFactory.create(new ParserModel(new FileInputStream(model_file)));
	System.err.println("done");
}
 

import opennlp.tools.parser.ParserFactory; //导入依赖的package包/类
@Override
public double scorePassage(Phrase q, Answer a, Passage p) {
	
	int countOfQuestionNPsInPassage = 0;
	try {
		//prep NLP tools
		if (!this.modelsAreInitialized) init();
		Parser parser = ParserFactory.create(this.parserModel, 20, 0.95);

		//create question parse
		Parse[] questionParse = ParserTool.parseLine(q.text, parser, 1);

		//create passage parses (one for each sentence)
		String[] passageSentences = this.DivideIntoSentences(p);
		Parse[] passageParses = new Parse[passageSentences.length];
		Parse[] tempParse;
		for (int i=0; i < passageSentences.length; i++) {
			tempParse = ParserTool.parseLine(passageSentences[i], parser, 1);
			passageParses[i] = tempParse[0];
		}
		
		//retrieve NPs from the question parse
		navigateTree(questionParse, 0, questionNPs);

		//retrieve NPs from the passage parse
		for (int i=0; i < passageParses.length; i++) {
			navigateTree(passageParses, i, passageNPs);				
		}
		
		//count the number of question NPs that are in the passage NP set (A)
		for (String qNP: questionNPs) {
			for (String pNP: passageNPs) {
				//System.out.println("comparing " + qNP + " with " + pNP);
				if (qNP.equals(pNP)) {
					//System.out.println("match found");
					countOfQuestionNPsInPassage++;
				}
			}
		}
		//System.out.println(countOfQuestionNPsInPassage);
		
		//count the number of all NPs that are in the passage NP set (B)
		//passageNPs.size();
		
	} catch (InvalidFormatException e) {
		e.printStackTrace();
	}

	//calculate A/B and return as the score
	//System.out.print("******** score:  " + (double)countOfQuestionNPsInPassage/passageNPs.size() + "  *******");
	//System.out.println(" count:  " + passageNPs.size() + "  *******");
	if (passageNPs.size() == 0)
		return 0;
	else
		return (double)countOfQuestionNPsInPassage/passageNPs.size();
}